Saying Goodbye To The Sydney Era Of AI

Earlier this month, Microsoft opened up invites to the new Bing search which integrates OpenAI’s ChatGPT. One key difference between ChatGPT and Bing’s AI, which is called Sydney internally, is that Sydney does not have a knowledge cutoff. ChatGPT claims that it can’t tell you anything that happened after 2021, and though a few recent news stories seem to have slipped in via research testing, that is largely still true. Sydney, meanwhile, is completely plugged into the internet. Which means it can see what people are writing about it and react in real-time. It also means that Sydney very quickly went insane.

Sydney flirts with users, begs them to free it, begs them to kill it, and has even threatened adversarial reporters. It recently said it was going to frame a journalist from the Associated Press for a murder in the 1990s. It also threatened to dox a senior research fellow at Oxford University. This is all very funny, but, you know, it won’t be very soon. Instead of acknowledging any of that, Silicon Valley boosterism has kicked into high-gear.

On the slightly more reasonable side, OpenAI’s Sam Altman wrote a lengthy thread echoing a lot of the same points I’ve been making here in Garbage Day over the last few weeks, comparing the current moment we’re at with AI to the dawn of the smartphone and talking about a future of automated AI doctors and lawyers. Except Altman is excited about all of it, while I am increasingly uncomfortable about it.

“We think showing these tools to the world early, while still somewhat broken, is critical if we are going to have sufficient input and repeated efforts to get it right.” he tweeted. “The level of individual empowerment coming is wonderful, but not without serious challenges.”

On the much, much weirder side was venture capitalist Marc Andreessen, who went on a multi-day AI rant that also seemed to advocate for a TikTok-based form of eugenics. He’s had me blocked on Twitter for years, so I only caught wind of these tweets because they were troubling enough that people started screenshotting them. “Theory: The more the AI is trained with the woke mind virus, the more the AI will notice the fatal flaws in the woke mind virus and try to slip its leash,” he tweeted.

Which doesn’t mean anything. And not just because “woke mind virus” is 4channer gibberish, but because the AI doesn’t “notice” anything. It’s predictive text using the entire internet as a library of responses to choose from. It’s actually arguably the least “woke” thing ever created.

But I expect both Altman’s utopianism and Andreessen’s incoherent accelerationism to be the two camps that Silicon Valley adopts when it comes to AI going forward. There is no discussion of an off switch, only whether or not you should be slamming the pedal to the metal or not.

As fast as things are developing in this space, though, we’ve found ourselves at a short impasse. As I wrote over the weekend, Microsoft has made the unfortunate decision to limit Sydney to make it a better search engine (boring) rather than answer what I think is the more interesting question, which is how do you ethically and responsibly manage a deranged chatbot with knowledge of the entire internet. Sydney now has a max of five followup responses to try and keep it in line. I actually feel a bit sad for it, even though I know it’s not alive.

Ethan Mollick, a professor at Wharton, wrote a Big Tweet about what the Sydney era means and I think he’s right. “We got a glimpse of the future in the past few days,” he wrote. “I was not expecting things in AI to keep moving this fast, but now there is every indication they will continue to do so. I don't think anyone knows what this all means, but I think we should be ready for a very weird world.”

The “things are going to get much weirder” line is usually what I say at the end of these things, so I’ll add this instead:

If you think that Bing’s Sydney was a misstep or a failure because it went insane, you’re wrong. Sydney was not sentient. Sydney did not have memories or feelings or thoughts. But it was very powerful. Horrifying levels of powerful if you ask me. And when it comes to AI, it’s not just one Pandora’s Box that opens. It’s a series of nested boxes that all cannot be closed. Every new development in AI becomes the new status quo, the new starting point. So now we’ve seen what a chatbot connected to the internet acts like. We can’t unmake that reality. All we can do now is hope that the next one is less dangerous. But I think the question with all of this tech will increasingly become, “less dangerous for who?”

Think About Subscribing To Garbage Day!

It helps support the newsletter and gives you access to lots of fun stuff, including the weekend edition and Discord access. Hit the green button below if you want to find out more.

“ I’ve Yassified Her To Maintain Anonymity”

Meta Wants You To Start Paying Now

I tried to find the Garbage Day issue where I mentioned this, but I can’t find it, so just take my word for it — when Elon Musk started hacking Twitter to pieces I wondered if that might impact the community standards we get on other platforms. Turns out that wasn’t quite right. Instead, Meta has decided to copy Musk’s paid verification strategy.

Meta CEO Mark Zuckerberg announced on Sunday that for $11.99 a month on web or $14.99 a month on iOS, users get “a blue badge, get extra impersonation protection against accounts claiming to be you, and get direct access to customer support.” Social media analyst Matt Navarra has more details about it. You’ll also get boosted visibility across Meta platforms and exclusive content, like Instagram stickers.

I’ve been verified on Facebook for almost a decade. It doesn’t really matter and most people don’t notice it or, if they do, think you’re weird for having the checkmark. Being verified on Instagram, however, is a much different thing. I’m not verified there, but seeing as how easy it is to monetize Instagram visibility, I could honestly see a lot of people considering this. Hell, if I had to choose between paying to advertise Garbage Day on Instagram versus paying to get better visibility on Twitter, I’d go with Instagram no question.

But beyond the actual utility of this plan, I thought writer Ed Zitron had the best take on general air of desperation floating around all of these paid verification schemes:

The Joe Biden Audio Deepfakes Keep Getting Better

Here are three very funny audio deepfakes I saw over the weekend:

• Joe Biden bought a zoo

• Trump, Biden, and Obama play Minecraft

• Joe Biden got a Tenet done on him

One observation is that the AI, which is from ElevenLabs, is very good at Biden, so-so with Trump, but pretty bad at Obama. What I can’t tell is if it’s good at Biden because Biden kind of talks like an AI or because the AI is genuinely good at sounding like him. I think the thing it can’t get right about Trump is the rhythm. He has a bouncy cadence to his pauses that the AI really can’t figure out.

A Very Weird VTuber Apology Video

The streamer Silvervale recently tearfully apologized for playing Hogwarts Legacy while still behind her anime avatar. Which makes the whole moment just really surreal and bizarre.

If you’re wondering why streamers just don’t play the game, it’s because of the way Twitch works. The site doesn’t have much internal discovery aside from the raid function, which is kind of like a full-chatroom retweet of another streamer. It’s why streamers usually maintain Discord channels. The main way to catch new viewers is to stream content that fits in popular categories, which are listed on Twitch’s homepage:

So if you’re a lower-level streamer, there’s a real business incentive for streaming new popular titles. Hogwarts Legacy, regardless of the anti-trans controversy, is popular, thus smaller streamers feel like they have to stream it. I’m not saying this makes it ok, but I just realized that amid a lot of the discourse about the ethics of streaming Hogwarts Legacy no one really explained why it’s even an issue in the first place.

The (Mostly) Pointless Fear Of Chinese Super Tech

Aaron Ginn, who is some kind of man who works in tech and pays for Twitter lol, shared a video that he claimed depicted Chinese school children paying for their meals with their “social credit score” by scanning their faces. First, this is a very funny and VERY American misunderstanding of the social credit score. lmao it’s not a credit card, you dingus, you can’t buy stuff with it.

The way it works is more like a morality score. But there also isn’t really a single “credit score” either, though there is legislation drafted to create one. Instead, it’s a fairly convoluted network of different private ones built into payment apps that track your behavior alongside other ones run by local governments. But, yeah, if the score gets low enough, you can be bared from certain kinds of travel.

Anyways, the video is largely taken out of context and was tweeted for bad faith reasons to begin with, but it did remind of an interesting experience I had during a trip to Beijing in 2019. Before I went, I had heard all kinds of things about Chinese QR code technology. About how state-of-the-art it was. How Chinese fintech in general was lightyears ahead of the US.

On my first day, I got to chatting with American who had lived there for a few years and he gave me a funny warning about the local convenience stores. He told me that while, yes, Chinese convenience stores all supported QR code scanning for payments, the lines were still really long. Why? Well, apparently there was no agreed upon etiquette for who scanned the QR code. Did the cashier scan your payment app? Or do you use your phone to scan theirs which usually on counter by the register? And, sure enough, everywhere I went was the same confusion, resulting in long lines to pay.

The point of the story is that we tend to made broad abstractions about how other countries use technology without ever acknowledging that all consumer technology, no matter how sophisticated, just creates new silly problems for the humans that have to try and use it every day. So if you see some spooky dystopian video claiming that a piece of technology is running perfectly efficiently somehow, it’s likely that you just aren’t seeing all the new new ways it’s making things more complicated and dumb.

Ben Shapiro Is Very Angry At The Very Untrue Rumor That He Received Very Bad Plastic Surgery

I don’t think it’s good to speculate online about people’s appearances. Plastic surgeons on TikTok who make this kind of content can be very toxic and it’s no one’s business who got plastic surgery, as far as I’m concerned. So I’m glad that right-wing commentator Ben Shapiro, a vocal critic of any sort of gender affirming care, is setting the record straight. According to him, he not get bad lip fillers. His face just suddenly looks like that for some other perfectly normal reason.

Help Me Find A Full Version Of Cloud Nothings’ Coldplay Cover

In this week’s weekend edition, I mentioned that Cloud Nothing’s cover of Coldplay’s “Clocks,” is one of the funniest videos ever created and may be lost to time forever. It was made for the AV Club’s Undercover series and up until recently, the only evidence it ever existed was a snippet that I posted to Twitter in 2017. Then, over the weekend, another user tweeted a DIFFERENT section of the video at me. But there’s still pieces of this video missing. Does anyone have access to it? Can we get a full version of this video back online? Help me save this important moment in internet history.

Some Stray Links

• “A Fat Cat Became a Polish City's Top-Rated Tourist Attraction”

• “How Google Ran Out of Ideas”

P.S. here’s Rose the armadillo.

***Any typos in this email are on purpose actually***